Since the discovery of the gene mutation responsible for FA, major advances have been made in our understanding of this disease, despite the difficulties encountered in gaining access to affected tissue in humans, especially the brain. In particular, FA can now be regarded as a true mitochondrial disease similar to a number of mitochondrial ataxias due to different genetic mechanisms [
38]. The respective roles of deficient respiratory chain function (decreased ATP synthesis) and impaired oxygen handling in the clinical course remain to be determined in most of these diseases. However, in the case of FA, impaired oxygen handling appears to be crucial in the cascade of events determining the onset of the symptoms and fits the progressive degenerative course of the disease. Obviously, emphasizing the role of hypersensitivity to oxygen in triggering FA does not shed any light on the undisputed contributory role of frataxin in ISC synthesis and/or stability [
39,
40]. Yet, available data indicate that both processes can be dissociated under a number of conditions in different models, including humans. Unfortunately, the molecular mechanism linking frataxin function to this hypersensitivity to oxygen is not yet established, similarly to the exact role of the protein in ISC synthesis, which is still a matter of intense debate. Frataxin has been claimed to activate import of iron into the cell [
41] and to chaperone iron [
42,
43] or ISC [
44,
45] to act as a partner for [
46] or inhibitor of [
47] ISC synthesis. How frataxin shortage in mitochondria results in impaired signalling of antioxidant defences in the cell cytosol has yet to be elucidated. So far, it is only known that frataxin deficiency in human cells interferes with the redox status of the cell [
21], thus presumably impairing the function of the Nrf2 [
9] and possibly PGC1α [
48] transcription factors.
Hypersensitivity to oxidant insult is a consistent feature of human cells [
5,
6,
8,
9] and animal models [
49,
50] with low frataxin content (Table
1). This observation should not be confused with an increased production of oxidative species and peroxidised products, yet has frequently been observed and reported (in about 100 papers), which either may not be observed, especially if respiratory chain activity is too severely depressed, or may not be accumulated if produced at low levels. Finally, in FA patients, oxidative markers and/or antioxidant enzymes are also modified in response to frataxin depletion [
17,
51‐
54], which was found to be an incentive to trial antioxidant molecules in this disease [
55‐
57].
Several ongoing clinical trials are evaluating interventions that target various steps in the pathogenic process: impaired frataxin synthesis, oxidative insults and iron accumulation [
58]. Idebenone, a short-chain coenzyme Q10 homologue, has initially been reported to be effective (at a dosage of 5 to 20 mg/kg/day) in preventing cardiac hypertrophy in most patients [
59], while having (at this dosage) little or no effect on the neurological abnormalities [
19]. Interestingly, while studying endomyocardial biopsies from a young patient with FA before and after idebenone treatment, we found that the drug largely restored the activity of ISPs [
60]. Since idebenone is a potent antioxidant, this finding suggests that loss of ISP activity
in vivo is due chiefly to increased oxidative degradation, and not to impaired synthesis, of these proteins. In this context, we may wonder why antioxidant therapies (for example, idebenone, coenzyme Q10) have such a limited impact on the neurological disease expression and/or course in FA. Indeed, recent widely based studies carried out in Europe (the MICONOS (Mitochondrial Protection with Idebenone in Cardiac Or Neurological Outcome Study) trial, comprising 232 patients with 162 on idebenone, mostly adults, at 13 centres; Andrews WT unpublished communication from the FARA meeting in Strasbourg, France, June 2011) and in the United States (the IONIA (Idebenone Effects on Neurological International Cooperative Ataxia Rating Scale Assessments) trial; 70 patients) [
61] did not confirm the previous reports of the beneficial effects of idebenone in this disease. However, as mentioned by the initiators of the MICONOS trial themselves, the results of the available studies, even if not statistically significant, still indicate a trend toward a positive effect of idebenone (as compared to placebo). A possibly underestimated factor that might prevent the results of such studies from reaching statistical significance in their relatively small cohorts is the occurrence of responsive and unresponsive patients reported in early FA patients treated with idebenone [
59], as has been true for a number of other mitochondrial diseases [
62]. Conversely, the negative conclusion of these latter trials, mainly carried out with adults, might be related to the age of the patients. Accordingly, even more recently this year, an open-label extension of the IONIA study (IONIA-E trial; 68 patients) concluded that idebenone may offer a therapeutic benefit to paediatric FA patients by stabilizing overall neurological function and improving fine motor skills and speech [
63]. There is no definite answer to this question; however, a number of indications suggest that intervention might be much too late. Indeed, it is a frequent observation in neurological diseases caused by gene mutations (nuclear or mitochondrial) encoding mitochondrial proteins that symptoms are subsequent to extended auto-amplifying cell death resulting from mitochondrial dysfunction rather than from mitochondrial dysfunction itself (for example, ATP decrease, metabolic blockade). This can be observed in a number of animal models where early, partial or tissue-specific inactivation of such genes (
Tfam and
Aif) results in a delayed neurological phenotype despite early mitochondrial dysfunction [
64,
65]. Thus, Tfam-depleted neurons, despite severe respiratory chain deficiency, are viable for one month in the mouse without showing signs of the neurodegeneration which precedes neurological symptoms [
64]. Similarly, despite early detectable complex I deficiency in the brain, the
Harlequin mouse with depleted Aif protein only manifests significant symptoms after several weeks or months of life in most individuals [
65]. Likewise, frataxin gene loss of function in FA, although it occurs early in embryogenesis, has a neurological impact several years later. Thus, the onset of neurological symptoms associated with impaired mitochondrial function might follow the loss of neurons rather than mitochondrial dysfunction
per se. Accordingly, any therapy aimed at counteracting mitochondrial dysfunction, regardless of the strategy used (modulating gene expression, gene therapy or pharmacological therapy) would be best tested if it preceded disease initiation.